rico Posted January 13, 2020 Report Share Posted January 13, 2020 1 minute ago, IU Scott said: Like I have said why do everyone has to over analyze every little aspect of the game. Without looking at any of this data and just going by what I have seen I could probably get most of the rankings correct. You analyze people analyzing it? What's the diff? Link to comment Share on other sites More sharing options...
FKIM01 Posted January 13, 2020 Report Share Posted January 13, 2020 Just now, Zlinedavid said: Because cumulatively to this point in the season, in the context of his model, they are. The last 5 games haven't been all that great, but that's what I appreciate about models like this. They're not reflective of the past week or two weeks. They're not based on opinions that are influenced by recent events. They take everything that has happened to date and push the outlook forward. That's apparently a difficult concept to grasp, but your attempt is appreciated. Link to comment Share on other sites More sharing options...
Hoosierhoopster Posted January 13, 2020 Report Share Posted January 13, 2020 KenPom's system is a model, a model for predictability of a team's success over the season. It's interesting, and useful, in that regard. It doesn't have to be perfect, and it's not the end all in assessing a team. It doesn't determine which teams get into the tourney. It's probably more useful for the gambler than guys like me, who just look at his "ranking" (whatever you want to call it) out of interest and thinking about how things may pan out. Link to comment Share on other sites More sharing options...
Zlinedavid Posted January 13, 2020 Report Share Posted January 13, 2020 1 minute ago, IU Scott said: What is that, we should allow bunch of computer geeks who probably don't know the game tell us who are the best teams. Given that I am one of those geeks, albeit in a different aspect/industry, yer damn skippy you should. 😂 I'll be the first to tell you, any predictive model or tool I've ever made is far from perfect. Compared with guys like Sagarin or Pomeroy, I'm not even in the ballpark. In the parking lot, maybe. It's a best guess, but at least it's a calculated guess vs a blind one. Link to comment Share on other sites More sharing options...
CincyHoosier Posted January 13, 2020 Report Share Posted January 13, 2020 3 minutes ago, IU Scott said: What is that, we should allow bunch of computer geeks who probably don't know the game tell us who are the best teams. You can look at it this way but that's not its purpose. It's a predictive model. It doesn't necessarily tell you who is playing the best basketball over the last week but it will tell you who is most likely to win the most games over the entire year. Link to comment Share on other sites More sharing options...
rico Posted January 13, 2020 Report Share Posted January 13, 2020 At the end of the day the only thing that matters is Selection Sunday. Link to comment Share on other sites More sharing options...
Proud2BAHoosier Posted January 13, 2020 Report Share Posted January 13, 2020 10 minutes ago, rico said: At the end of the day the only thing that matters is Selection Sunday. Amen and Amen..... Link to comment Share on other sites More sharing options...
Zlinedavid Posted January 13, 2020 Report Share Posted January 13, 2020 9 minutes ago, IU Scott said: Like I have said why do everyone has to over analyze every little aspect of the game. And the answer to this...which will probably be the equivalent of dividing by zero for you....is because I legitimately enjoy it. Analytics like this don't detract from my enjoyment of the game. They enhance it. They give me hints on areas to watch for as I'm watching the games. Is this team supposed to be a good defensive team? If so, why did we put 85 up on them? Is it a change in the trend for them? Is the model wrong? Did we find a loophole? Why? It's how I see the world. I'm a process oriented thinker. Parts of a whole, cause and effects....it's subconscious to me. That's how I'm wired. That's how I get enjoyment out of watching. And yes, there are other whackjobs like me that do the same. The ironic thing is that everything you hate about analytics is the same improper way that people in sports use them. "This model says X, so I'm going to do X." Anyone in any type of organization that takes the word of a model as gospel without any type of question is a fool. Hell, I always question my own models, and if nobody else does, I start to worry. They are predictions and indicators. If my model shows that Metric X is going to be 15% low, that's a prediction. Don't bet the farm on it being 15%. Now, what I will say with confidence is that if I said it would be 15% down and it wound up being 12% or 17% down.....that's a damned victory. Was I exact? No, but I may have caught the start of a trend that could be reversed with the right action that I may have overlooked otherwise. It pointed to the right area, in the right direction and at the approximate amount. That's what it's supposed to do. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 15 minutes ago, Zlinedavid said: And the answer to this...which will probably be the equivalent of dividing by zero for you....is because I legitimately enjoy it. Analytics like this don't detract from my enjoyment of the game. They enhance it. They give me hints on areas to watch for as I'm watching the games. Is this team supposed to be a good defensive team? If so, why did we put 85 up on them? Is it a change in the trend for them? Is the model wrong? Did we find a loophole? Why? It's how I see the world. I'm a process oriented thinker. Parts of a whole, cause and effects....it's subconscious to me. That's how I'm wired. That's how I get enjoyment out of watching. And yes, there are other whackjobs like me that do the same. The ironic thing is that everything you hate about analytics is the same improper way that people in sports use them. "This model says X, so I'm going to do X." Anyone in any type of organization that takes the word of a model as gospel without any type of question is a fool. Hell, I always question my own models, and if nobody else does, I start to worry. They are predictions and indicators. If my model shows that Metric X is going to be 15% low, that's a prediction. Don't bet the farm on it being 15%. Now, what I will say with confidence is that if I said it would be 15% down and it wound up being 12% or 17% down.....that's a damned victory. Was I exact? No, but I may have caught the start of a trend that could be reversed with the right action that I may have overlooked otherwise. It pointed to the right area, in the right direction and at the approximate amount. That's what it's supposed to do. Most of this is not directed at people on this board but just the fact that about everything we see in sports is based on analytics. It is worse in baseball than any other sport and everything they think is based on analytics. i go to the Reds board, don't post but just read and when discussing a playerthere opinion on tha tplayer is only by their analytics. it is always based on what their WAR is or soemthing like that and not your actual opinion on that player. You watch some of the shows on MLB network and it is always a fight between the anlytics people against the old school fpormer players or managers. When it comes to the tournament I rather have people put in place where there only job is to watch games and be on the selection committee. I just feel who ever comes up with these formulas put things in them that they feel are the most important things and other might not think so. They also can put things into the formulas that will further their viewpoints and maybe not what is best for the overall game. Link to comment Share on other sites More sharing options...
rico Posted January 13, 2020 Report Share Posted January 13, 2020 3 minutes ago, IU Scott said: Most of this is not directed at people on this board but just the fact that about everything we se ein sports is based on analytics. It is worse in baseball than nay other sport and everything they think is based on analytics. i go to the Reds board, don't post but just read and when discussing a playerthere opinion on tha tplayer is only by their analytics. it is always based on what their WAR is or soemthing like that and not your actual opinion on that player. You watch some of the shows on MLB network and it is always a fight between the anlytics people against the old school fpormer players or managers. How many times have we been through this before? Too many with you. But FT% is an analytic. As is FG% and so on and so forth. Scott, you have dealt with them your whole life without realizing it. The thing you struggle with now is the advanced metrics, I can speculate as to why but I won't. In this day and age it is broken down just a little bit further. That is a good thing. But I would also say that numbers don't lie, but they don't tell the whole story either. If it was just metrics and everything went "according to plan" then Vegas wouldn't exist. Link to comment Share on other sites More sharing options...
GaloisGroupe Posted January 13, 2020 Report Share Posted January 13, 2020 KenPom is just a statistical model. Like all statistical model, users must to determine how to properly interpret and use it (or ignore it). This requires understanding topics that were previously addressed (e.g. the use of last years data). There is certainly an association between a team having a better KenPom ranking and having better odds of winning a game (after appropriate adjustments for home vs. road, etc.). And, where the interest lies, it occasionally will provide us with unexpected, but correct, relationships (e.g. predict the rise of St. Mary's). PS. In defense of using previous years data to seed early season models: Teams that are good in year n tend to be good in year n + 1 (similar things can be stated for bad teams). With UNC etc., this season seems to be a bit of an outlier. Until better information reveals itself, one might as well use this. A lot of us worry about the team going 1 - 11 in a 12 game stretch. Our implicit assumption is that there is an association between last years performance (and team make-up and weaknesses) and this years performance. "But one thing I hope to convince you of in this book is that data are profoundly dumb." - Judea Pearl, The Book of Why Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 8 minutes ago, rico said: How many times have we been through this before? Too many with you. But FT% is an analytic. As is FG% and so on and so forth. Scott, you have dealt with them your whole life without realizing it. The thing you struggle with now is the advanced metrics, I can speculate as to why but I won't. In this day and age it is broken down just a little bit further. That is a good thing. But I would also say that numbers don't lie, but they don't tell the whole story either. If it was just metrics and everything went "according to plan" then Vegas wouldn't exist. The reason I don't like them is because I think they are hurting the product we see on the field or court. In basketball analytics tells the people that you should only shoot layups and 3 pointers and anything else is a bad shot. I totally disagree with that and with taking away the middle game it makes it easier to defend and that is part of the reason we have see the offensive struggle we see today. In baseball all the analytics tells us that pitchers should not pitch over 100 pitches in a game and it leads to the over use of the bull pen and causing to many pitching changes.\ yes I know stats has been used for years and I still care about some of them but these advance metrics are not needed in MY OPINION. Also I couldn't care less about Vegas because I don't gamble and I watch sports because I love the game and watch it for the actual game and nothing else. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 Rico what are you confused about because you asked why I did not like analytics and I told you because I feel it hurts the product we see on the court or field. Link to comment Share on other sites More sharing options...
Zlinedavid Posted January 13, 2020 Report Share Posted January 13, 2020 20 minutes ago, IU Scott said: When it comes to the tournament I rather have people put in place where there only job is to watch games and be on the selection committee. I just feel who ever comes up with these formulas put things in them that they feel are the most important things and other might not think so. They also can put things into the formulas that will further their viewpoints and maybe not what is best for the overall game. Want to know what the dream outcome of anyone that builds statistical models is? For the model to exactly match reality, 100%. Unless you have the ability to bend reality to fit your personal agenda, your model will be dismissed as inaccurate. Want to know what I would rather have? A committee that knows how to use statistics, but also balance them out with their own empirical observations and opinions. Use a model to arrive at a tie, and then use your opinions as a tiebreaker, or vice versa. Use them as they were intended. Not as substitutes for human discretion, but to enhance and refine human discretion. Link to comment Share on other sites More sharing options...
rico Posted January 13, 2020 Report Share Posted January 13, 2020 5 minutes ago, IU Scott said: The reason I don't like them is because I think they are hurting the product we see on the field or court. In basketball analytics tells the people that you should only shoot layups and 3 pointers and anything else is a bad shot. I totally disagree with that and with taking away the middle game it makes it easier to defend and that is part of the reason we have see the offensive struggle we see today. In baseball all the analytics tells us that pitchers should not pitch over 100 pitches in a game and it leads to the over use of the bull pen and causing to many pitching changes.\ yes I know stats has been used for years and I still care about some of them but these advance metrics are not needed in MY OPINION. You do realize that the Selection Committee uses analytics? They have their formula...you realize that right? Link to comment Share on other sites More sharing options...
rico Posted January 13, 2020 Report Share Posted January 13, 2020 2 minutes ago, IU Scott said: Rico what are you confused about because you asked why I did not like analytics and I told you because I feel it hurts the product we see on the court or field. Your incessant bitching. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 3 minutes ago, rico said: You do realize that the Selection Committee uses analytics? They have their formula...you realize that right? Yes I do and would rather them use them only as a tie breaker rather them picking teams solely on them. I would rather the selection committee use their own brain and opinion on who they think are the best teams. Link to comment Share on other sites More sharing options...
GaloisGroupe Posted January 13, 2020 Report Share Posted January 13, 2020 2 minutes ago, Zlinedavid said: Want to know what the dream outcome of anyone that builds statistical models is? For the model to exactly match reality, 100%. Unless you have the ability to bend reality to fit your personal agenda, your model will be dismissed as inaccurate. Want to know what I would rather have? A committee that knows how to use statistics, but also balance them out with their own empirical observations and opinions. Use a model to arrive at a tie, and then use your opinions as a tiebreaker, or vice versa. Use them as they were intended. Not as substitutes for human discretion, but to enhance and refine human discretion. The goal of statistical modelling is the second bullet: provide additional information. Any inferential statisticians who try for the first bullet should be barred from practice! I think a lot of complaints about advanced analytics and modelling is the mis-use of advanced analytics and modelling. They are just additional information. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 4 minutes ago, rico said: Your incessant bitching. So it is alright for you to come on here and state your opinions on things but I can't. got it Link to comment Share on other sites More sharing options...
rico Posted January 13, 2020 Report Share Posted January 13, 2020 1 minute ago, IU Scott said: So it is alright for you to come on here and state your opinions on things but I can't. got it Whatever my friend. Stick your head in the sand all you want. Link to comment Share on other sites More sharing options...
zerawkid Posted January 13, 2020 Report Share Posted January 13, 2020 7 minutes ago, rico said: Your incessant bitching. There was probably a better way to say that. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 Another thing about the selection committee using analytics I think it hurts the really good mid major program who has a great record. The analytics will tell you a team who finishes 8-12 in the big ten is better and more deserving than a 28-3 mid major team and I don't think it is right. Link to comment Share on other sites More sharing options...
IU Scott Posted January 13, 2020 Report Share Posted January 13, 2020 1 minute ago, zerawkid said: There was probably a better way to say that. He is good, I have no problem with it. Link to comment Share on other sites More sharing options...
Zlinedavid Posted January 13, 2020 Report Share Posted January 13, 2020 8 minutes ago, IU Scott said: In basketball analytics tells the people that you should only shoot layups and 3 pointers and anything else is a bad shot. This is a perfect example of a sweeping generality being misused. As an overall trend, they are correct, and I can mathematically prove it to you. Keep in mind, none of these first few examples will be realistic, but to help you see the objective. Let's assume that in a 100 possession game and all other factors being the same, I'm 50% likely to make a 3 point shot, 50% likely to make a midrange jumper and 50% likely to make a layup. Which should I take more of? Easy, the three point shot. My likelihood of making them are equal, and in the end, I'd score 150 points vs 100. Now, let's adjust that to 40% for a 3 point shot, 50% for a midrange jumper and 60% for a layup. Again, all other things being equal, in 100 possessions, what should I do? The answer is either the 3 or the layup. Why? Because over 100 possessions, your expected points would be 120 for either of those vs 100 for the midrange jumper. Now, what I'm about to say may be contradictory to you. I believe the concept above is 100% mathematically sound. I also believe that any coach that uses the above as gospel and does not look into it further is an idiot. A team's roster isn't taken into account. The opponent isn't taken into account. The opponent's trends (yay..more stats!) aren't taken into account. Maybe a team's defensive scheme defends the arc very well, and maybe they have a great shot blocker. In this case, the likelihood of a zero point possession for a 3 attempt or a layup are higher than they normally would be. So at this point, you're very much correct that the midrange would be effective. Now, to what extent should I use it.....depends on my team. Catch a team off balance and sagging back, then use the threat of the established midrange to open the three point line back up. I didn't lose sight of the overall principle, but I used all aspects of the game available to set the game up to be statistically in my favor. This is the step that gets missed. Link to comment Share on other sites More sharing options...
FKIM01 Posted January 13, 2020 Report Share Posted January 13, 2020 6 minutes ago, zerawkid said: There was probably a better way to say that. ...but probably not a funnier one. Even Scott laughed. I trust it was mostly said tongue in cheek with just a dash of annoyance. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.